The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx -y to download the @mjpitz/mcp-rfc package from the npm registry at runtime during the execution of RFC lookup commands.
[REMOTE_CODE_EXECUTION]: The skill executes code from a third-party package (@mjpitz/mcp-rfc) that is not listed as a trusted vendor. This execution occurs via npx when the agent attempts to fetch or search for RFCs.
[COMMAND_EXECUTION]: The skill invokes shell commands using python3 $MCP_CALL to bridge communication with the RFC MCP server, providing a local execution path for the downloaded package.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection by ingesting document content from the external IETF RFC registry.
Ingestion points: Full RFC text and specific sections are retrieved via get_rfc and get_rfc_section calls within SKILL.md.
Boundary markers: The instructions do not define delimiters or special tokens to separate RFC content from the agent's internal system instructions.
Capability inventory: The skill allows the execution of system commands through the $MCP_CALL environment variable and npx.
Sanitization: There is no evidence of sanitization or filtering of the RFC content before it is processed by the agent.

rfc-lookup